Sounds like the beginning of a 21st century horror movie: Your browser history has been public all along, and you had no idea. That’s exactly how you feel now in the new offline app Meta AI, where many people are posting their supposedly private conversations with a chatbot.
When you ask a question to the artificial intelligence, you have the option to click the “Share” button, which redirects you to a screen with a preview of the message, which you can then publish. But some users seem blissfully unaware that they are publicly sharing these text conversations, audio clips, and images with the world.
In the Meta AI app, people have asked for help with tax evasion, to find out if their family members will be arrested for their involvement in criminal activities, or how to write a letter of recommendation for an employee who is in trouble with the law, including the person’s name and surname. Others, such as security expert Rachel Toback, have found examples of people’s home addresses and confidential court records, among other private information.
With every passing second, these seemingly innocuous queries in the Meta AI app are getting closer to viral chaos. In a matter of hours, more and more posts appeared on the app that clearly indicated trolling, such as someone sharing their resume and asking for a job in cybersecurity, and someone with a Pepe the Frog avatar asking how to make a bong out of a water bottle.
If Meta wanted people to actually use its Meta AI app, then public embarrassment is certainly one way to get attention.
